haiku.rag/haiku_rag_slim/haiku/rag/client
Yiorgis Gozadinos 30123585ed
Break cross-database RRF rank ties by retrieval score
Disjoint corpora give every database's rank-r candidate the same RRF
score, and the stable sort resolved those ties to lancedb.databases
declaration order, discarding the retrieval scores entirely. Ties now
break on the raw retrieval score, which is uncalibrated across indexes
but only ever orders candidates within one rank tier: the databases in
a fusion share an embedder and ran the same search type, and it can
never lift a candidate above another rank. Hybrid per-database scores
are themselves rank-derived, so exact agreement still ties and keeps
configured order, deterministically. The n > limit depth quota is
unchanged, pending the retrieval eval.
2026-09-01 13:47:54 +03:00
..
__init__.py Finish the comment pass, and escape document fields everywhere Rich renders 2026-08-28 15:34:47 +03:00
agents.py Name the databases before the model runs 2026-08-27 17:07:41 +03:00
documents.py Name the database a write wrote to 2026-08-28 08:47:36 +03:00
downloads.py Delete comments that restate the line below them 2026-08-20 15:22:33 +03:00
exceptions.py typed UnsupportedSourceError replaces pipeline string-marker matching 2026-05-26 11:44:46 +03:00
processing.py Resolve file:// URIs to paths through url2pathname 2026-08-21 10:22:26 +03:00
rebuild.py Repair an FTS index that covers no rows on write 2026-08-31 18:54:48 +03:00
scope.py Finish the comment pass, and escape document fields everywhere Rich renders 2026-08-28 15:34:47 +03:00
search.py Break cross-database RRF rank ties by retrieval score 2026-09-01 13:47:54 +03:00
session.py Finish the comment pass, and escape document fields everywhere Rich renders 2026-08-28 15:34:47 +03:00
titles.py Extract title generation into client/titles.py 2026-04-24 14:42:53 +03:00